Your AI agents' memory, hosted in France
Give your AI agents a structured memory: OpenViking unifies context, resources, and skills through a file system paradigm. DINAO hosts your instance — your contexts remain on our French servers.
What is OpenViking?
OpenViking is an open-source context base designed specifically for AI agents. It unifies the context management an agent needs — memory, resources, and skills — through a file system paradigm, enabling hierarchical context delivery and self-evolving memory.
Instead of flat vector storage, OpenViking organizes everything in a hierarchical virtual file system: memories, resources, and skills are stored in directories accessible via the viking:// protocol. A three-level structure (L0/L1/L2) loads content on demand, significantly reducing inference costs.
The platform automatically manages sessions: it compresses content and updates the user's and agent's memory directories. Ready-to-use Docker images facilitate deployment with persistent state. It is an ideal building block to equip your agents with rich, structured memory deployable with full sovereignty.
Host OpenViking at DINAO
Resource tiers compatible with OpenViking prerequisites (minimum 1 vCPU / 1 Go / 5 Go). Hosted in France, fully managed.
- 1 dedicated vCPU
- 2 Go RAM
- 20 GB NVMe
- Daily backups
- Managed & monitored by DINAO
- 2 dedicated vCPU
- 4 Go RAM
- 40 GB NVMe
- Daily backups
- Managed & monitored by DINAO
- 4 dedicated vCPU
- 8 Go RAM
- 80 GB NVMe
- Daily backups
- Managed & monitored by DINAO
- 8 dedicated vCPU
- 16 Go RAM
- 160 GB NVMe
- Daily backups
- Managed & monitored by DINAO
This application uses AI
The container hosts the application, not the AI engine (which requires dedicated GPUs) : inference is handled externally, with your own provider key. Prioritize a sovereign engine — Mistral AI, NumSpot, Scaleway, or OVHcloud AI Endpoints (France, GDPR) ; an international provider (OpenAI, Anthropic…) only if a specific capability requires it. Inference subscriptions are not included in hosting.
Technical details
You might be wondering…
What is OpenViking exactly for?
OpenViking is a context base for AI agents: it stores and organizes their memory, resources, and skills as a hierarchical file system, making the context more structured and reusable than with simple vector storage.
How is it different from a standard vector database?
Instead of flat vector storage, OpenViking offers a tree structure (L0/L1/L2) accessible via the viking:// protocol, with on-demand loading. Relevant context is delivered at the right time, reducing inference costs.
Does my data go to an AI provider?
It depends on the connected LLM. In local mode (open-weight on your instance), everything stays in France. With an external API, the transmitted context passes through the provider — see the sovereignty note.
Where is the data hosted?
On DINAO infrastructure in France, in one of the available data centers. Agent memories and contexts remain on your dedicated instance.
Who manages updates and security?
DINAO. We install versions, apply patches, monitor the instance, and perform backups. You develop your agents, we operate the platform.
